Researched LLM comparison metrics and 15 leading benchmarks, leveraging academia to build an LLM evaluation website on Streamlit for guiding model selection and comparison in Vodafone’s GenAI community.
Co-creating a computer vision-based airport security system using Raspberry Pi, 5G technology, and edge computing on AWS.
Implementing a dockerized RT-DETRv2 object detection model for real-time object and crowd monitoring.
Working in a team to develop a patent-pending ML solution utilising Vodafone's signal data for nowcasting fog and humidity, contributing to data exploration and cleansing efforts with Pandas.
Collaborated with team to develop speech-to-text translation prototype using Gemini 1.5 model on Streamlit platform.
Assessed text quality, fluency, safety, and verbosity of six generative language models (Gemini and OpenAI) using Vertex AI’s GenAI Evaluation Service, presenting insights to wider teams.
Data and AI Analyst
Vodafone
Newbury
09.2023 - 09.2024
Co-created a Product Information Chatbot POC, leveraging RAG and GPT-3.5, and reducing review time from 45 minutes to seconds through an end-to-end ingestion pipeline, designed to update the RAG datastore with newly updated documents, significantly enhancing efficiency.
Designed a comprehensive Power BI dashboard to manage FTE projections over 12 months for 550 employees, with a contract value of 281.3 million. Secured strong buy-in from four external teams, who now rely on it during monthly demand planning meetings.
Facilitated onboarding process by establishing groundwork needed to support external GenAI vendor partnership.
Junior Business Analyst
Change Grow Live
Leeds
11.2022 - 09.2023
Performed UX research and requirements gathering from a pool of 6,290 employees for software development and IT initiatives, supporting the UK’s largest drug and alcohol electronic healthcare record system.
Facilitated focus groups to identify user pain points.
Coordinated with the development team to resolve software issues.
Designed comprehensive UI mockups and UX artifacts using software like Miro, Balsamiq, and Visio.
Applied Agile methodologies, prioritizing the sprint backlog using impact and priority-type labels.
Education
Master of Science - Business Analytics
ESADE Business & Law School Ramon Llull University
Barcelona
10-2022
Bachelor of Arts - International Tourism Management
Leeds Beckett University
Leeds
07-2021
Skills
Programming Languages: Python (intermediate), R (basic)